30 September 2026 in Environmental safety and natural resources
Purpose: to develop an automated data orchestration framework that consolidates heterogeneous biotechnological datasets from disparate laboratory sources into a unified structured relational knowledge base, serving as a scalable foundation for artificial intelligence model training in organic waste management.Design/methodology/approach: a modular Python/Flask-based software framework was developed, incorporating fuzzy-logic column-mapping algorithms, a dynamic SQL generation engine, and a real-time validation hint mechanism. The system processes XLSX and CSV files, transforming them into unified relational schemas with user-adjustable semantic mappings.Findings: the system successfully consolidates data from both batch and continuous anaerobic digestion regimes. The column_mapper module resolves cross-laboratory nomenclature conflicts, while the sql_generator ensures relational data integrity. A systematic review of over 50 studies (2010–2025) confirmed that the median dataset size in AD research is merely 353 data points; the proposed framework overcomes this barrier by automating cross-laboratory data aggregation.Theoretical value: the study establishes the theoretical concept of a "digital common denominator" for cross-laboratory biotechnological analysis, grounded in fuzzy set theory and formal ontological semantics.Practical implications: the framework provides a ready-to-use tool for transforming isolated laboratory spreadsheets into structured deep-learning datasets with cross-experimental anomaly detection capabilities. The validation hint system prevents low-quality data from entering the relational database.Originality/value: for the first time, an automated semantic orchestration concept for heterogeneous biotechnological data is proposed, combining fuzzy-logic mapping and dynamic SQL synthesis – a qualitative advance over traditional manual preprocessing and isolated data silos.Limitations/future research: expansion of supported data formats; cloud platform integration; development of automated cross-experimental anomaly detection modules; training and validation of AI models on the resulting unified datasets.
10 November 2025 in Матеріали міжнародної науково-практиченої конференції "Екологія. Людина. Суспільство"
Anaerobic digestion (AD) is a process that has found increasing relevance for sustainable waste management and bioenergy production. Nevertheless, microbial interactions and process dynamics make it challenging to maximize and predict biogas production. Recent breakthroughs in artificial intelligence (AI), particularly in artificial neural networks (ANN), have provided new opportunities for solving these problems and drastically enhanced process optimization, predictive capability, as well as perpetual stability. This paper aims to provide a review of recent literature covering the application of AI, specifically Artificial Neural Networks (ANN), integrated with various optimization algorithms, including Genetic Algorithms (GA) and Particle Swarm Optimization (PSO), in anaerobic digestion systems from 2022 to 2024. The results of the evaluation demonstrate that ANN models exhibit a significant improvement over conventional kinetic-based approaches, characterized by lower prediction errors, higher biogas production, and greater robustness.